Live data from Hacker News

Google top 1000 sites: Interesting stats about them

blog.sucuri.net

11–16 of 16 posts

Re: Google top 1000 sites: Interesting stats about them

#12
post #10

Earlier quoted context omitted.

How do they even tell, given that sites can avoid having .php/.asp/.jsp in their URLs if they want to?

We checked the extensions, the "Server:" option in the header and the "X-Powered-by" option. We tried our best:)

So were the other 68.7% all unknown?

Re: Google top 1000 sites: Interesting stats about them

#13
post #2

Nginx with 14% of market share is indeed very interesting, very close to IIS (17%).

The stats can be misleading, nginx is very good at being a reverse proxy or software load balancer and tends to be put to use in those contexts with pass-thru to existing web servers.

Because the stats look at headers, the last header before hitting the internet will be the nginx caches.

Re: Google top 1000 sites: Interesting stats about them

#14
post #7

How can they detect the programming language in use other than by looking at .php, .aspx, .jsp, etc? You won't see this on a most professionally-authored sites that use a router and RESTful URLs.

That's not what RESTful means. What your URLs look like has absolutely nothing to do with REST -- the whole point is that it treats URLs as opaque references to other similarly hypertextual resources.

Re: Google top 1000 sites: Interesting stats about them

#16

I'd like to see sites that use mootools, Prototype, YUI, Dojo, etc. Those are some simple statistics to compile I'd think (filename based, or just simple regex of the first 200 characters).

http://wappalyzer.com/stats/cat/JavaScript%20frameworks (click the links for more detail).
Post reply on HN